Conventional video coders calculate a processing amount (an amount of coding processing) necessary for coding video data in coding the video data stored in a frame memory and perform processing for reducing the calculated processing amount. Specifically, the amount of coding processing is reduced by simplifying moving prediction below a slice layer or a macro block layer in forward reference image coding or bidirectional reference coding. Such processing reduces an amount of coding processing so that coding of video data is completed within time calculated based on a frame rate (see, for example, Japanese Patent Application Laid Open Publication No. 2005-57750A).
In the conventional video coders, however, because an amount of processing for coding video data is reduced prior to execution of the coding, new video data may be overwritten in the frame memory to cause image noise when a real coding time period overruns predetermined time (for example, time calculated based on a frame rate). When an amount of coding processing overflows, error notification is output to cause interruption of the processing, resulting in temporally discontinuous coding. Thus, the coding system may be broken.